RICARDO'S WORK AS VIEWED BY LATER ECONOMISTS

by

Kenneth J. Arrow D TICTechnical Report No. 531 ELECTE

August, 1988 NOVO 91988

PREPARED UNDER THEUNITED STATES OFFICE OF NAVAL RESEARCH

ONR-NO0014-86-K-0216

L tbd Lbuthoa Unlmited

THE ECONOMICS SERIES

INSTITUTE FOR MATHEMATICAL STUDIES IN THE SOCIAL RCIENCESFourth Floor, Encina Hall

Stanford UniversityStanford, California

94305

RICARDO'S WORK AS VIEWED BY LATER ECONOMISTS*

by

Kenneth J. Arrow**

/ David Ricardo was a peaceful man, well liked and admired for both his

intellectual and his personal traits by his colleagues and rivals whether on

the Stock Exchange, in the nascent field of political economy, or among the

members of Parliament. He could maintain personal friendship and well-

behaved exchange of ideas with someone as strongly removed in both religion

and economic doctrine as Thomas Malthus. The intellectual strength of his

written work could dominate the thought of such a great mind as that of John

Stuart Mill and rouse thp writer Thomas de Quincey from his opium-riddled

state to renewed mental vigor.

Yet Ricardo's posthumous reputation has been very variable. Many econom-

ists in the decades after his death attacked the realism of his theories, par-

ticularly with regard to population pressure.i Samuel Bailey did not think

much of his logical powers. John Stuart Mill took him as supreme authority,

yet almost all Mill's original contributions tended to subvert Ricardo's doc-

trines and move them in the direction of neoclassical theory. Karl Marx, like

Mill, put Ricardo in the highest rank, yet used his theories for directions copyNiSPECTO

totally antithetical to any view Ricardo might have held. The neoclassical 6

*This work was supported by Office of Naval Research Grant N00014-86-K-

0216 at the Institute for Mathematical Studies in the Social Sciences,Stanford University, Stanford, California.

This paper originally appeared in German in, K. J. Arrow, M. Ricardo, andH. C. Recktenwald, David Ricardo: Eine moderne Wurdigung, in the series,Vademecum zu einem frUhen K--tssiker, DUsseldorf: Verlag Wirtschaft undFinanzen, 1988. -

*Department of Economics, Stanford University, Stanford, California.

-" "

-2-

theory arose as a reaction to the limitations of Ricardian classicism, and its

supporters can never have a wholly approving tone; at best, the most favorable

statements they can make are that Ricardo in part anticipated them, most

notably in the marginalist character of his theory of rent. The most famous

neoclassical judgment, if put more strongly than most would, was that of

Jevons [: "that able but wrong-headed man, David Ricardo, shunted the car of

Economic science on to a wrong line." (The figure of speech is an interesting

anachronism, since the railroad was still in its infancy at Ricardo's death.)

The general neoclassical view of Ricardo as at best an imperfect and

none-too-coherent predecessor has continued to the recent past; nor is it con-

fined to the more mathematical wing. Frank Knight, greatly influential in the

formation of the Chicago School, severely criticized Ricardo's failure to con-

sider economics as a process of resource allocation. Joseph Schumpeter,

magisterially surveying economic analysis from a self-chosen viewpoint above

all schools, hardly has a good word for Ricardo's methodology. The "Ricardian

vice" to Schumpeter is that of assuming enough variables as given outside the

system that the remaining analysis becomes trivial.

There is one striking exception to the generally negative neoclassical

view of Ricardo, that of Alfred Marshall. Marshall's repeated admiration for

Ricardo, was expressed at greatest length in Appendix I of the Principles.

LIKe Schumoeter, Marshall saw himself in a very superior position in the whole

world of economics, and even towards his hero, Ric~fdo, there is more than a

touch of condescension. Ricardo is given to an excessive love of abstraction,

no doubt, says Marshall, the result of Ricardo's "Semitic" heritage. (The

-3-

great biological developments of the nineteenth century had their dark side in

the use of science to justify racism, and this example is far from the only

one in Marshall.) But a candid reading of Marshall makes his admiration hard

to understand. Marshall's theory is at the bottom no different from Jevons's

or Walras's, though of course enriched in many ways and given more concrete

development. Demand and its determination through utility play a central

role, totally absent from Ricardo. There is of course emphasis on the cost of

production as a determinant of value, but Jevons understood its role equally

well, and Walras was as clear as could be desired on the symmetry between cost

and utility. The most distinctive aspects of Marshall's doctrines, the notion

of external economies to reconcile perfect competition with increasing returns

and the development of the role of time in equilibrium ("long-run" versus

"short-run") are precisely those most remote from Ricardian thought. Most of

Marshall's Appendix I consists of showing that Ricardo was "feeling his way"

towards modern and "correct" ideas, such as the distinction between marginal

and total utility.

There is one curious episode which may illuminate or obscure Marshall's

views on Ricardo. Marshall reviewed Jevons's Principles of Political Economy

for a magazine called, Academy, in 1872. Instead of recognizing and applaud-

ing the novelty of the ideas or criticizing them from, let us say, Ricardian

or Millian lines, the review can only be described as carping. From the ana-

lytic viewpoint, the main criticism was that Jevons had not gone far enough in

taking a truly general equilibrium viewpoint. After Marshall's death, a copy

of the review was found with a note appended. Its flavor can best be present-

-4-

ed in some direct quotations R: "I looked with great excitement for Jevons'

Theory; but he gave me no help in my difficulties and I was vexed. I have

since learned to estimate him better. ...[H]e seemed perversely to twist his

own doctrines so as to make them appear more inconsistent with...Ricardo's

than they really were. But the genius which enabled Ricardo...to tread his

way safely through the most slippery paths of mathematical reasoning, though

he had no aid from mathematical training, had made him one of my heroes; and

my youthful loyalty to him boiled over when I read Jevons' Theory. The editor

of the Academy, having heard that I had been working on the same lines, asked

me to review the book: and, though a quarter of a century has passed, I have a

vivid memory of the angry phrases which would force themselves into my draft,

only to be cut out and then reappear in another form a little later on, and

then to be cut out again."

Clearly, Marshall had already developed at least the basic lines of his

theory when Jevons' book appeared; the editor of the Academy knew that. Mar-

shall freely acknowledged his indebtedness to the neoclassicists avant la let-

tre, J. von Thunen and A.A. Cournot. Is it too much to conjecture that he was

disappointed or worse to find himself unexpectedly anticipated by Jevons?

Perhaps the defense of Ricardo could justify to Marshall an anger about prior-

ity unacknowledgeable even to himself.

Since I have made one conjecture, I might as well make another. The

supreme authority of Marshall at Cambridge well beyond his active career is

not in dispute. The modern use of Ricardo as a bludgeon against neoclassical

-5-

economics is certainly centered at Cambridge, with some alliances among Itali-

an economists; is it possible that it derives at least the aura of respectabi-

lity and legitimacy from Marshall's emotional pro-Ricardo bias?

At any rate, the next great step in the history of the reception of

Ricardian thought is the designation of Piero Sraffa as the editor of the Roy-

al Economic Society's edition of the collected works of Ricardo. That Ricardo

should be so singled out by the Royal Economic Society hardly needs explana-

tion. Ricardo's influence on the history of economic thought is great, whe-

ther one thinks it good or bad. Keynes's dominance of the Royal Economic

Society continued the still uncontested Cambridge leadership created by Mar-

shall, and Keynes correctly foresaw that Piero Sraffa would be a superb edi-

tor. No doubt the choice of Sraffa was the more logical since Sraffa disliked

lecturing and therefore had the more leisure for the task.

Sraffa had his own intellectual agenda. He did not like the subjective

elements in neoclassical theory and therefore wanted to play down the import-

ance of demand. Above all, he was clearly influenced by Marx's version of

Ricardo, so in particular he refused to accept the position that the supply of

capital was the result of subjective decisions about the timing of consump-

tion. Obviously, there is a major ideological position here, one which Marx

understood well enough when he ridiculed Senior's argument that saving is due

to "abstinence." The rich, exemplified for Marx by the Rothschilds, are hard-

ly "abstaining." Sraffa began in the late 1920's the elaboration of his theory

of "production of commodities by means of commodities," a work that was to see

-6-

publication only in; 1960 . But through the oral traditions of Cambridge,

Sraffa's version of Ricardo was to influence strongly the work of Joan Robin-

son and of such younger Italian and British economists as L. Pasinetti, N.

Garegnani, A. Roncaglia, and J. Eatwell.

While Sraffa was diligently and very slowly completing his edition of

Ricardo (for which, in the end, he needed the collaboration of M.H. Dobb)

and the exposition of his own doctrines, Cambridge and the economic world was

rocked by a another revolution, which drew to it most of the bright young

spirits, including Joan Robinson. Keynes was, in successive steps, developing

the new attack on past economics which culminated in The General Theory.

Keynes shared with Sraffa a lack of interest in marginal economics and especi-

ally in the marginal utility theory of demand; he did not share Sraffa's ideo-

logical presuppositions or his appreciation of Marx. From Keynes's viewpoint,

there was little distinction between classical and neoclassical economics;

both mistakenly assumed full employment, Say's Law, and the impossibility of

general gluts. Ricardo as well as A.C. Pigou had to be purged. Keynes did

not confuse an antiquarian interest in seeing Sraffa produce a good edition of

Ricardo with any excessive respect for an obstacle to what he saw as correct

thinking. His understanding of past economists was primarily dictated by

their consistency with his current thought. There could be little sympathy

between Sraffa and Keynes, much as Keynes admired Sraffa's intellect. As Joan

Robinson says, "Sraffa had shown a draft [of Production of Commodities by

Means of Commodities] in 1928. Keynes evidently did not make much of it and

Sraffa, in turn, never made much of the General Theory.'14

-7-

It is therefore surprising that a wave identified as Ricardian or, at

least, "neo-Ricardian," emerged in the shadow of Keynes after his death. Joan

Robinson sought to remain true to both traditions in some way, by dropping

Ricardo's assumption of full employment. Others, like Sraffa himself, seemed

to be indifferent to the full employment hypothesis, though one would have

thought a position on this question to be basic to the formulation of a

coherent model. If prices do not have the property that all markets clear,

then there must be a hypothesis that the price on a non-clearing market may,

for some one reason, remain unaffected.

What is common to all varieties of neo-Ricardian thought is a generally

antagonistic attitude to capitalism and a tendency to interpret the history of

economic thought and in particular the emergence of neoclassical thought as an

attempt to divert attention from Marxist critiques and from the allegedly sub-

versive character of Ricardian thought. The neoclassicists, it is alleged,

are the more or less conscious apologists for capitalism. Ricardo's doctrines

are taken to have two implications which later economists, both his immediate

successors and the later marginalists, were concerned to deny: that labor is

the source of all value and that profits have to be thought of as a deduction

from output and therefore from labor's share.

It is also sometimes pointed out that Ricardo's doctrine of rent and in

particular of the tendency of population to rise to the agricultural carrying

capacity is pessimistic about the prospects of workers under capitalism. It

is argued that, as soon as workers become politically significant, it is

-8-

necessary to show them that capitalism does not imply their permanent misery.

However, the scarcity of land is pessimistic with regard to any social system,

not merely capitalism; indeed, Malthus introduced his principle of population

precisely to argue that social reform of any kind cannot produce continuous

progress, as Condorcet and Goodwin had argued. It is no wonder therefore that

the possibility of diminishing returns was dropped by Marx and equally by

Sraffa and Joan Robinson.

It would be too far from the aim of this essay to examine whether or to

what extent the criticism of Ricardo by his successors was motivated, con-

sciously or unconsciously, by the fear of giving aid to the enemies of capi-

talism. Like any historical inquiry, it would have to be conterfactual, and

it is not very clear what would count as relevant evidence.

Instead let me make a few remarks about the fundamental character of

Ricardo's system, as I see it. I am a practicing theorist, not a scholar in

the history of economic thought, and my reactions are no doubt influenced by

my own views. Indeed, it is hard to write intellectual theory and, for that

matter, any kind of history, without framing the past through the perspective

of the present. It can justly be remarked about history that it is as reveal-

ing about the time when it was written as about the time which is its ostens-

ible subject matter. Nevertheless, it is morally incumbent on us to try as

far as possible to put ourselves in a position contemporary with that of our

author.

-9-

Was Ricardo asking a different set of questions about the economy than we

would today? The mainstream today asks about the allocation of resources:

what incentives direct resources into one activity or another, into one kind

of consumption as opposed to another, or into investment for future production

as against consumption today, to what extent and under what conditions are

resources directed efficiently (with some appropriate definition of efficien-

cy). After Keynes, we might add the question whether resources are all uti-

lized or are there some left idle, a question which, from a certain point of

view, is again about a rather gross form of inefficiency. In principle, we

ask about allocation among individuals or among owners of different factors of

production, but it must be recognized that distributional questions are not

asked very loudly or answered very well.

In a competitive system, resource allocation is in particular directed by

prices, and so a determination of prices becomes a central issue. Prices of

course include both wages and rates of profit.

Adam Smith indeed asked what governed the wealth of nations. But Ricardo

apparently asks a different question. The Preface of The Principles of Poli-

tical Economy and Taxation opens boldly, "The produce of the earth ...is

divided among three classes of the community; namely, the proprietor of the

land, the owner of the stock or capital necessary for its cultivation, and the

labourers by whose industry it is cultivated. ...To determine the laws which

regulate this distribution, is the principal problem in Political Economy."

This then is a book on what was later called the "functional" distribution of

-10-

income. Indeed, there are chapters on wages, on rent (the understanding of

which Ricardo regarded as basic to the distribution of income), and on

profits.

But these chapters are preceded by a chapter on value and followed by

many chapters on the workings of markets, on foreign trade, on the incidence

of various kinds of taxes, and on other matters only vaguely associated with

the distribution of the produce of the earth (national income) among the dif-

ferent classes of society. Indeed, if Ricardo were to announce a determina-

tion to study the allocation of resources in the modern sense but with the

knowledge available to him, it would be hard to know how the topics treated

would be very much different. For example, labor is needed for production,

but the quantity of labor depends on the existence of output (agricultural and

other) to provide the standard of living desired by workers. This is governed

by the willingness of capitalists to save and thereby provide circulating

capital and by the availability of land to produce the food component of the

standard of living. The supply of land is not given, but expanding it

requires using lower qualities. At each stage, there is a complex interlock-

ing set of resource requirements to meet demands which in turn affect sup-

plies.

The main thrust of Ricardo's system is a bold attempt to determine values

independent of dem,-nd considerations. I do not mean that Ricardo thought of

the alternative of a role for demand in determining values and rejected it.

Rather he did not really conceive of this alternative. Clearly he lacked

that very elementary tool, the demand schedule. Neither he nor any of his

-11-

contemporaries had the explicit relation of demand to price, though not many

years were to pass before Cournot and J.S. Mill were to develop it independ-

ently. But I do not think, as some neo-Ricardians seem to, that there was in

any sense an intended repudiation of the demand schedule. Indeed, some of

Ricardo's analysis can only be made sensible on the basis of such a concept.

Thus, if there is more capital in an industry than is needed to meet the

demand, it is asserted that the market price will fall, so that the capital

will be earning less than the normal rate of return and therefore will exit.

Evidently, the price falls so that demand will rise to use the excess capaci-

ty.

Without the language of demand schedules (and supply schedules for fac-

tors), there is really no way of opposing a theory that demand influences

prices to a theory where prices are determined purely by technology. But

Ricardo works hard at stating the assumptions needed to achieve a purely

supply-based pricing theory. Normal wages are exogenous in that they repre-

sent conventional demands for subsistence. They are not physiological, and

they do include manufactured goods as well as food; but they are enforced by

the willingness of laborers to reproduce themselves. Manufacturing operates

at constant returns to scale. As we now understand the matter, it is import-

ant to assume the absence of joint products. Rent is handled along well-known

lines. Capital by and large, though there are some exceptions, is circulating

capital, an advance of the wage-goods during the period of production. If,

then, the period of production is the same in all industries, manufactured

goods exchange in proportion to the labor embodied (directly or indirectly),

-12-

and they also exchange on the same basis with agricultural product on the mar-

ginal land. The analysis of agricultural product is indeed in accordance with

modern marginalist principles, except for the fixed proportions between capit-

al and labor.

But the slightest variation in this sketch destroys the simplicity of the

system, in particular, the determination of normal prices within a system

which excludes demand. I do not intend to engage in the process of scoring

Ricardo for his adequacy in anticipating subsequent developments in economic

theory. But some deficiencies should have been clear. First of all, there

are, as Ricardo himself states briefly, many kinds of labor with very differ-

ent wage rates. This fact alone should make the meaning of a labor theory of

value obscure. Ricardo, like Marx after him, contents himself with taken

these differences as given. Smith's environmental determinism implied that

all individuals are basically alike; occupation is a decision variable, so

that net advantages, rather than wages, are equalized. The most cursory know-

ledge of the world would suffice to show many wage differences are not

explained by net advantages. But Ricardo does not offer even that hypothesis.

Second, the period of production does differ from one industry to anoth-

er, as Ricardo emphasizes at considerable length. He does try to minimize the

practical extent to which relative prices would be affected, but he does not

deny the theoretical impact. Logically, for a given willingness to save, how-

ever defined, the rate of profit will depend on the relation between the

demands for goods with long and short periods of production (I am clumsily

-13-

trying to avoid the controversial term, "capital-intensity").

Third, competing uses of the same land for different crops or for urban

versus agricultural purposes would, as J.S. Mill later recognized, create the

same indeterminacy in the system of price determination. One can understand

how the dazzling beauty of a simple model of value determination would blind

the eyes of any investigator to these problems, however obvious they might

seem.

Even Ricardo's most famous accomplishment, the law of comparative advan-

tage in foreign trade, is incomplete, though not wrong, without demand consid-

erations, as J.S. Mill was the first to observe. Ricardo can only determine

'imits on price ratios.

The difficulty in incorporating demand into Ricardo's system was, no

doubt, connected with the problematic relation between use value and exchange

value. Smith had already argued that usefulness cannot explain value, because

water is clearly more valuable than diamonds. Ricardo therefore concluded, as

did Marx after him, that use value was necessary for exchange value but did

not explain its magnitude. Jeremy Bentham with his usual clarity of thought

did in fact resolve the paradox by distinguishing between marginal and total

utility; but as usual with him the fragment was not published. Bentham was

very close to James Mill, Ricardo's good friend, and cooperated with Ricardo

for a while in an attempt to create a school along Benthamite lines, but some-

how never conveyed this insight to either or to his amanuensis, J.S. Mill.

-14-

The lack of a satisfactory theory of utility was damaging in another

way. It prevented a coherent statement for defending the policies which

Ricardo favored, especially free trade. Since foreign trade is an exchange of

equal values, there is no gain in terms of value. Ricardo is reduced to vague

phrases like, "increasing the general mass of productions," in explaining the

benefits of trade.

What Ricardo did contribute was, for good or bad and probably some of

both, the style of economi, inalvsis that has dominated the science, abstract

reasoning, proceeding from a few .-- A principles that commend themselves to

the reader as reasonable to a multitude of conclusions checked at points

against everyday observation. Today, indeed, we have added more systematic

use of data and sophisticated empirical analysis to casual observation. The

avoidance of inconsistent and ad hoc reasoning is made into a major virtue.

It was Ricardo more than anyone else who created the flavor of economic theory

and analysis, as much in the neo-Ricardians as in the neoclassicists.

-15-

References

W. S. Jevons, The Theory of Political Economy, Fifth Edition. New York:

Augustus M. Kelley, 1957. Preface to the Second Edition, p. li.

A.C. Pigou (editor), Memorials of Alfred Marshall. London: Macmillan,

1925. Pp. 99-100.

P. Sraffa, Production of Commodities by Means of Commodities. Cambridge:

Cambridge University Press, 1963.

J. Robinson, "Keynes and Ricardo." Reprinted in J.C. Wood (ed.), David

Ricardo: Critical Assessments, Volume IV. London and Sydney: Croom

Helm. P. 148.